Blindsight by Peter Watts. It's a near-future first contact story where a crew of meta-humans, led by an AI-enhanced autistic space vampire, try to respond to an alien ship that appears in Jupiter's orbit and scans the earth. It goes deep on neuropsychology and philosophy while the crew struggles to make sense of the event.

It has some extreme weirdness but it's all scientifically justified. The book has copious endnotes explaining the basis for everything from blindsight to Cotard's delusion to the vampirism and does a great job bridging the gap between bizarre reality and pure fiction.

Not really. This article does a decent job of discussing the topic. In short, it is theoretically possible that certain cancers could be "transmissible," but the subject isn't very well understood. However, a big study of medical registry data in Scandinavia found no statistically significant increased risk of cancer among people who had received a blood transfusion from someone who developed cancer.

Key takeaway is here:

>The researchers identified 978 cases of cancer among all the blood recipients but after statistical analysis they found no excess risk of cancer overall among individuals who had received one or more blood products from a precancerous blood donor. The relative risk was not substantially affected by sex age, calendar period, or number of transfusions. What is more, there was no excess risk when patients who received blood from people with cancers at sites that are thought to have the highest risk of metastasising through blood---the lung, liver, skeleton, and central nervous system---were combined.

Similar studies in the US have found the same thing: "Results did not imply any concrete association between cancer risk and history of blood transfusion. These findings would help in debunking the myth of increased cancer risk following blood transfusion."